C HAPTER 2 Plant–Plant Interactions in Tropical Agriculture Luis García-Barrios CONTENTS Introduction Interference Intraspecific Interference in Monocrops Evaluating Intraspecific and Interspecific Interference Effects in Two-Species Stands Interference, Coexistence, and Overyielding in a Two-Species Stand: The Competitive Production Principle Revisited A Word on Interference in Weed–Crop Systems Studying Competitive Interactions in Many Species Stands Allelopathy Facilitation Indirect Facilitation The Interplay of Plant Interactions in Environmental Gradients: Some Consequences for Tropical Agriculture Plant Interactions and System Design and Management Frequency of Land-Use Rotation Intercropping Intensity Relay Crops Crops Associated with Service Plants Full Intercrops The Percentage of Tree Canopy Cover in the Field © 2003 by CRC Press LLC Direct Facilitation Interference in Productivity Gradients Conclusions References INTRODUCTION At the end of the last millennium, tropical agricultural systems extended over a surface of about 20 million square kilometers, occupied by nearly a billion people from 50 different African, Asian, and Latin American underdeveloped countries (Anders, 1990). Tropical agricultural systems (TASs) are found in an enormous variety of contrasting environmental conditions (MacArthur, 1976a; National Research Council, 1993): topography ranges from flat lowlands to very steep high- lands; soils range from moderately fertile to very unfertile; temperature ranges from cool to hot; humidity ranges from extremely wet to very dry; primary vegetation ranges from tropical rainforests to semiarid shrub lands. The variety of TASs is even greater (ranging from shifting agriculture to highly technified agroindustrial planta- tions) due to the distinct social, cultural, and economic conditions and to the different intensities under which these environments and their resources are being used (Ruthenberg, 1976; Hougthon, 1994). Nevertheless, most of tropical agriculture shares the following relevant features: • Their environmental conditions are frequently restrictive and fragile. Most tropical soils have low fertility due to water erosion, leaching, and acidification in more humid climates and to high temperatures and wind soil erosion in dryer ones (Bennema, 1977; Lal and Miller, 1990). Warm humid and subhumid regions are exposed to explosive pest populations, flooding, and crop storage and transporta- tion problems (National Research Council, 1993), while semiarid regions suffer frequently from prolonged droughts (Okigbo, 1990). • Most TASs are part of the livelihood of peasant smallholders who often confront severe socioeconomic restrictions on production (Okigbo, 1990). A significantly smaller proportion of TASs are part of large agroindustrial plantations, which in some cases can extend over vast portions of land in the subhumid and humid tropics (National Research Council, 1993). • Many TASs are established in the so-called megadiversity regions of the world, and all TASs contain and foster significantly more biodiversity than their corre- sponding temperate counterparts (Perfecto et al., 1996; Collins and Qualset, 1999). • Market-based economies and social pressure on land are rapidly driving TASs toward increasing levels of intensification, specialization, and simplification (Mac- Arthur, 1976b; Hougthon, 1994), which produce short-term economic benefits to farmers, but only add to these systems’ economic and ecological fragility in the long run (Vandermeer et al., 1998). Traditional TASs were able to persist during millennia in spite of many restric- tions. Land-use intensification and other expressions of social change are confronting them with enormous sustainability challenges: in the absence of significant soil conservation measures and costly exogenous inputs, agricultural productivity rapidly decreases and potential land degradation increases (Ruthenberg, 1976; Lal and Miller, 1990). Semiarid tropical regions have always had significantly higher © 2003 by CRC Press LLC restrictions than more humid ones (Stewart, 1990). Unfortunately, as population and land-use intensification increase, the natural productivity and potential land degra- dation are dangerously closing toward their lower ends (Figures 2.1–2.3). Figure 2.1 Schematic representation of the effects of climate and land-use intensification on agricultural productivity in the tropics, according to Ruthenberg (1976) and Mac- Arthur (1976b). Figure 2.2 Schematic diagram of potential land degradation in the tropics, in relation to climatic aridity. (Modified from Lal and Miller, 1990, Figure 2.) Humid Subhumid Semiarid calid semicalid calid Agricultural productivity Shifting agriculture Intensive agriculture 1.0 0.8 0.6 0.4 0.2 0 Potential land degradation Humid Subhumid Semiarid Arid Amazon and Congo basins Semi deciduous forest Derived savannas Grass savannas with intense burning Sahel With deforestation and intensive land use © 2003 by CRC Press LLC Sustainability has become a strongly debated concept and a major issue in international development policies (Schaller, 1993; Demo et al., 1999). It has focused attention on the need to develop and promote the necessary social conditions and the ecologically sound technologies required for a sustainable agriculture in the tropics (National Research Council, 1993; Hatfield and Keeney, 1994; Buck, Lassoie, and Fernandes, 1999). In order to increase ecological sustainability in these fragile environments, the ideal TAS should have at least the following biologically based attributes: • A high plant cover or residual biomass for efficient light and water capture, constant soil protection, and soil organic matter accumulation • A low dependence on costly and noxious external inputs accomplished by a relatively small harvest, by removal of nutrients in relation to total biomass, by efficient nutrient recycling, and by natural means of pest control • A variety of different types of crops and associated beneficial organisms as a means for increasing and diversifying produce and income, reducing the risk of total losses, and accomplishing the conditions established in the first two attributes In short, preserving and promoting plant diversity within TASs is an important (and maybe the most available) ingredient in the endeavor for a more sustainable agriculture in the tropics (Edwards, 1990; Altieri, 1992; Edwards et al., 1993; Vandermeer, 1995; Tilman, 1996; Tilman, Wedin, and Knops, 1996; Vandermeer et al., 1998; Thrupp, 1998). Of course, means to solve labor and capital constraints for maintaining, developing, and promoting such biodiverse systems are also required. The ecology, agronomy, and economy of multispecies systems have received increasing attention (Kass, 1978; Vandermeer, 1989). A number of advantages over Figure 2.3 Schematic diagram of potential land degradation in the tropics, in relation to population per arable hectare. (Modified from Lal and Miller, 1990, Figure 1.) 1.0 0.8 0.6 0.4 0.2 0 Potential land degradation 0 5 10 15 20 25 In uplands and harsh climates Mexico Tanzania India China Kenya Egypt Projected population per arable hectare by year 2000 © 2003 by CRC Press LLC monospecific crop systems have been repeatedly pointed out. In my opinion, such advantages are frequently overgeneralized in the sustainable agriculture outreach literature. It is necessary to stress that (1) not all potential crop mixtures can coexist, nor do they result in economically and ecologically sound systems; (2) their different benefits are not equally valued by all farmers, nor can they be maximized simultaneously; and (3) they demand more complex tasks and higher inputs of labor and of ecological knowledge than are sometimes available (García- Barrios et al., 2001). Multispecies systems have developed historically as a continuous trial-and-error process through which specific groups of farmers have identified crops and other associated organisms that can be brought together to their advantage. Social and environmental production conditions change continuously and, with them, the via- bility of seemingly well-established agricultural systems. In order to develop and maintain agrodiversity in the face of economic, social, and environmental changes, which nowadays tend to advance at a speedier pace than the farmer’s empirical exploration and adaptation capacity (García-Barrios and García-Barrios, 1992), it can be useful to support the farmer’s effort with a more systematic and extensive theoretical and practical exploration, which can help evaluate current multispecies systems and design those appropriate to new circumstances. From an ecological perspective, the emphasis should be on developing and applying knowledge and skills that can help farmers to manipulate ecological interactions within these systems and foster those interactions that enhance crop productivity and lower risk while reducing external inputs and conserving soil, water, and biological resources. During the last century, plant ecologists and agronomists have developed impor- tant ecological knowledge on plant interactions (i.e., interference, allelopathy, facil- itation) (Harper, 1990). Mainstream agronomic research has focused on monocrop situations, and its interest in the details of ecological interactions has been marginal. Plant ecology research has been more concerned with the theory and details of such interactions and has engaged in studying far more diverse and complex plant com- munities. Agroecologists are making important efforts in bringing together the con- tributions of both disciplines in order to help understand, develop, and successfully manage both simple and complex multispecies agricultural systems (e.g., De Wit, 1960; Vandermeer, 1981, 1989; Firbank and Watkinson, 1990; Radosevich and Rousch, 1990; Gleissman, 1998). The purpose of this chapter is to contribute to this effort. The following three sections briefly review the current ecological knowledge on plant interference, allelopathy, and facilitation. Interference has received much atten- tion in crop research and is therefore presented more directly related to agricultural systems and in an analytical fashion. Allelopathy and facilitation are far more diverse interactions and have seldom been studied analytically (but see Vandermeer, 1989). They are treated on a more general basis, derived from the plant ecology literature, but the implications for TASs are discussed. The section titled “The Interplay of Plant Interactions in Environmental Gradients” analyzes the interplay of these plant interactions and how they are modified in productivity gradients. This topic has recently attracted the attention of plant ecologists and, in my opinion, has important implications for TASs when considering how plant–plant interactions can vary in © 2003 by CRC Press LLC the heterogeneous and contrasting soil and climatic conditions encountered in the tropics at the regional, local, and field levels, and how these interactions can change as a consequence of land-use intensification. The section titled “Some Consequences for Tropical Agriculture” examines the way positive and negative interactions come together in the major TASs and the possibilities of benefiting from these interactions through proper management. For the purpose of this discussion, TASs are classified according to (1) the permanence of a specific plant assemblage on a patch of land or, conversely, the frequency of land-use rotation; (2) the intensity of intercropping, meaning the number, type, and level of spatiotemporal concurrency of crops within the field; and (3) the percentage of tree canopy cover in the system. Finally, the section titled “Plant Interactions and System Design and Management” presents some concluding remarks. INTERFERENCE In order to grow and develop, all plants require solar radiation, water, nutrients, and space. As a plant grows, it continuously expands the above- and below-ground zone of influence from which it can actually or potentially acquire such resources. Interference occurs when two plants that have developed overlapping zones of influence reduce one or more of these resources to the point where the growth, survival, or reproductive performance of at least one of them is negatively affected (Begon, Harper, and Townsend, 1986). Interference interactions between growing neighbors constitute a dynamic process whereby both individuals continuously mod- ify the other’s above- and below-ground environment and respond to such modifi- cations. Goldberg (1990) considers that a plant’s competitive ability comprises the capacity to affect environmental resources (effect competitive ability) and the capac- ity to tolerate reduced environmental resources (response competitive ability). The effect on resources is related to uptake traits as well as to nonuptake processes that affect resources either positively or negatively, while response to resources is related to the balance between rates of resource uptake and loss at the individual or popu- lation level. (See Goldberg [1990] Table 2 for a useful description of these traits and processes.) The ways in which the net effect of one species on another is determined by their effect on and response to environmental resources are nontrivial and are just beginning to be understood (Goldberg, 1990). Agricultural systems are commonly established at densities that imply highly competitive conditions. In multispecies agroecosystems where different crops, weeds, and trees grow together, the interplay between intraspecific and interspecific competitive abilities strongly influences what species will be able to coexist and what the per-species and per-stand yield will be. Understanding the mechanisms of competitive abilities for the sake of predicting community structure and productivity has proved elusive and controversial (Tilman, 1987; Grace, 1990). Further compli- cations arise because competitive abilities above and below ground are context dependent, for they change in complex ways along resource gradients. Nevertheless, important progress has been made on the subject; see Vandermeer (1989), Grace (1990), and Holmgren, Scheffer, and Huston (1997) for further details. © 2003 by CRC Press LLC I take a very general and phenomenological approach to interference by focusing mainly on how net intraspecific and interspecific interference can be evaluated and on their consequences for multicropping yields. I begin by looking at the effects of intraspecific interference at the stand level, both for the sake of analyzing its role in tropical monocrops and to better understand the interplay of both kinds of inter- ference in multiple cropping systems. Intraspecific Interference in Monocrops The probability that an individual plant will be adversely affected by its conspe- cific neighbors increases the more their zones of influence overlap, as a consequence of growth or increased plant density. Competitive effects in a dense monospecific stand have important consequences for the population as a whole. They strongly influence its size distribution dynamics (Koyama and Kira, 1956; Gates, 1982; Hara, 1988), its self-thinning trajectory (Westoby, 1984), and the particular form taken by the yield–density relation (Willey and Heath, 1969; Vandermeer, 1984b). Individual seedlings seldom grow at the same pace within a monospecific stand due to small differences in genotype, germination time, microenvironmental condi- tions, or tissue loss to herbivores and pathogens. These initial differences are further enhanced nonlinearly by competition, more so when it is intense and asymmetric (Thomas and Weiner, 1989; Weiner, 1990). This leads the approximately normal seedling size distribution (Figure 2.4a) to become increasingly skewed to the left as individuals grow (Figure 2.4b). Eventually, the smallest individuals, most strongly affected by interference, die out. Thus, if interference is sufficiently intense, density is reduced as the average plant in the population grows in size, and eventually Figure 2.4 Schematic representation of change in size distribution skewness in a monospe- cific plant stand. Subtle differences in time and size of birth as well as unequal intrinsic growth rates are further exaggerated through interference: (a) early stage; (b) late stage. (Modified from Begon and Mortimer, 1986, Figure 2.10.) Individual seedling dry weight (mg) 0 10 20 30 40 50 0 10 20 30 40 Individual plant dry weight (g) 10 20 30 40 50 0 10 20 30 40 (a) (b) Number of individuals Number of individuals © 2003 by CRC Press LLC stabilizes at a value that is specific to the particular environment and species. For any initial arbitrary density, the biomass of the average plant in the population grows to a critical point beyond which further increase can only be achieved with a concomitant loss of individuals (Begon and Mortimer, 1986) (Figure 2.5). The so- called self-thinning rule (Westoby, 1984) states that in an overcrowded situation, the number of individuals in the stand must be reduced tenfold in order for a survivor to increase its biomass a hundredfold. The most obvious effects of intraspecific interference are reduced growth rate, final biomass, and seed set weight of the average plant. These per-plant variables normally are descending geometric functions of sowing density. When the population as a whole is considered, above-ground plant biomass per unit area is commonly an asymptotic function of sowing density, while seed yield per unit area is either asymptotic or quadratic (Willey and Heath, 1969; García-Barrios and Kohashi, 1994; Figure 2.6). Asymptotic behavior occurs when reduction in per-plant growth is exactly compensated for by the increase in plant number, which leads to constant final yield, a condition most common in species with plastic, indeterminate growth. Quadratic behavior occurs when increased density disproportionately reduces seed setting or seed weight and when severe self-thinning cannot be compensated for by the remaining population. These conditions are mostly found in species with less plastic, determinate growth. Figure 2.5 Schematic diagram of the self-thinning process in a monospecific plant stand: the relation between plant density and mean individual’s weight. Maximum individual weight is marked with an asterisk. (Modified from Begon and Mortimer, 1986, Figure 2.13b.) Number of individuals per unit area 1 10 100 1000 10000 0.1 1 10 100 1000 Slope = ca. -3/2 * Individual dry weight (g) © 2003 by CRC Press LLC Evaluating Intraspecific and Interspecific Interference Effects in Two-Species Stands In multiple species stands, both intraspecific and interspecific interferences are encountered simultaneously. Comparing the intensities of intraspecific and interspe- cific interference helps to explain plant species coexistence, plant mixture overy- ielding, and weed–crop interactions. Consider the case where two monocrops (spe- cies A and B) are sown in separate unit-area plots, each at its optimum density (i.e., the density that produces maximum per-unit-area yield). In such conditions, each species’ population uses resources as efficiently as it can. Then consider a substitutive intercrop where 50% of plants in the B monocrop are substituted for by species A plants. When comparing the latter species’ per-plant yields in monocrop and inter- crop, six basic scenarios can result (Figure 2.7). Condition 3 in the figure is an interesting reference case, where A’s per-plant yield remains the same as in the monocrop. This suggests that per-individual interspecific and intraspecific interfer- ence should be equal (i.e., A and B individuals are competitively equivalent). In other terms, although intraspecific interference is obviously reduced in the intercrop due to substitution, it is exactly compensated for by interspecific interference. As shown in Figure 2.7a, interspecific interference can also be greater or lower than intraspecific interference. It can also be zero (no interspecific interference), or even negative if net interspecific facilitation occurs. The consequences on A’s per-unit- area yield are shown in Figure 2.7b. A similar analysis is presented in Figure 2.8 for an additive intercrop in which 50% of B’s optimum monocrop population is added to the full A monocrop. Again, condition 3 is the competitive equivalence case. Both analyses can also be done for species B. I will now briefly address the consequences of some of these possible outcomes for species coexistence and land-use efficiency in an intercrop. Figure 2.6 Schematic diagram of the parabolic relation between plant density and yield. w means average yield per plant and Y means yield per unit area. In the example, optimum density = 10; optimum w = 0.5; maximum Y = 5. 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Y w Number of plants per unit area Optimum y Optimum w 0 1 1.5 2 2.5 3 3.5 4 4.5 5 0.5 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 © 2003 by CRC Press LLC Interference, Coexistence, and Overyielding in a Two-Species Stand: The Competitive Production Principle Revisited For a multiple crop system to be viable, its component species must be able to coexist, and the system must have advantages over its competitors (the corresponding monocrops). The land equivalent ratio (LER) is the most commonly used intercrop performance index (Willey, 1979). LER is defined as (A’s intercrop yield/A’s maximum monocrop yield) + (B’s intercrop yield/B’s maximum monocrop yield) The use of this criterion assumes that the farmer is interested in both crops, and it defines how many monocrop surface units yield the same as one intercrop surface unit. An LER greater than 1.0 implies both coexistence and overyielding, while an LER less than 1.0 can still mean the former, but not the latter. Figure 2.7 Identifying the relative intensity of interspecific interference in a 50%A:50% sub- stitutive intercrop design: (a) representation on a per-plant basis; (b) representa- tion on a per-unit-area basis. Six possible scenarios are depicted in each figure. α AA = intraspecific interference between two A plants. α AB = interspecific interfer- ence exerted by plant B on plant A. Condition 3 is a reference case, where per- individual interspecific and intraspecific interferences are equal (α AA = α AB ). As in Figure 2.6, optimum density = 10; optimum w = 0.5; maximum Y = 5. See text for further details. w 1 2 3 5 4 6 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Y 1 2 3 5 4 6 0 number of sp. A plants per unit area (a) (b) 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 Six possible scenarios: 1: α AB >> α AA 2: α AB > α AA 3: α AB = α AA 4: α AB < α AA 5: α AB = 0 6: α AB < 0 © 2003 by CRC Press LLC [...]...w 1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0 .2 0.1 0 6 5 4 3 2 0 1 2 3 4 5 1 6 7 8 9 10 11 12 13 14 15 6 5 Six possible scenarios: 1: α 2: α 3: α AB AB AB 0 0 >> α AA > α 4 = α 3 AA AA 4: α AB < α 5: α AB = 0 AA 2 6: αAB < 0 0 1 2 3 4 5 6 7 8 1 9 10 11 12 13 14 15 5 4.5 4 3.5 3 2. 5 2 1.5 1 0.5 0 Y Number of sp A plants per unit area Figure 2. 8 Identifying the relative intensity of interspecific... M2 0.5 M2 0.75 M2 M2 Crop 2 yield per unit area Figure 2. 9 The intensity of interference and facilitation in a diculture define its position in a diculture evaluation plane Points A–F correspond to 1 + 1 additive designs; points G–I to 1:1 substitutive designs When monocrops 1 and 2 (at their respective optimum densities M1 and M2) are added, some possible outcomes are (A) 1 facilitates 2 while 2 strongly... further details, and Figures 2. 23 and 2. 24 for examples, which the reader can easily place in this figure © 20 03 by CRC Press LLC (1) (6) Biodiversity and structural complexity (2) (7) (3) (4) (5) Land use intensification (crop permanence) Figure 2. 23 As land use is intensified and rotation cycles are shortened or eliminated, the biodiversity and structural complexity of tropical agricultural systems... answered in terms of interference are (1) how strong is the per-individual interference (or per-unit-leaf-area effect; Kropff and Lotz, 19 92) of the weed over the crop, and (2) how much can interference be reduced by increasing crop density or hastening crop emergence? (Zimdahl, 1988; Kropff, Weaver, and Smits, 19 92; Liebman and Gallandt, 1997) Figure 2. 10 shows the typical effect of relative weed density... gathering impractical (Tilman, 1990) Following Goldberg and Werner (1983) and Tilman’s (1990, 1996) work with natural communities, I would © 20 03 by CRC Press LLC 1.0 b) Relative crop yield a) 0.75 0.50 0 .25 0.0 -2 0 0 20 40 Weed emergence delay (days) Figure 2. 10 0 25 50 75 Relative weed density (%) The relative timing of weed emergence (WED) can either revert or aggravate the negative effect of relative... affected by environmental productivity © 20 03 by CRC Press LLC low interaction between above- and below-ground interference at the ends of a water or nutrient gradient, and strong interaction in the middle Two hypothetical examples of this situation are presented in Figure 2. 18 For canopy-dominant species and subcanopy shade-tolerant species, one can expect below-ground competition either to decrease... be classified and placed accordingly in the resulting three-dimensional space Figures 2. 23 and 2. 24 show some major TASs, classified according to these criteria Frequency of Land-Use Rotation Most TASs derive originally from forest clearing (Hougthon, 1994) Under very low social pressure on land, shifting agriculture constitutes an adequate long-term rotation system where mature forest alternates with... ever-increasing interference on the other, while the latter reduces it interference on the former Such dominance-reduction relations between individuals can eventually lead the weaker competitor to exclusion In substitutive intercrops, the dominant © 20 03 by CRC Press LLC BC LER =2 Crop1 yield per unit area M1 F 0.75 M1 D LER=1 0.5 M1 Additive designs G Substitutive designs 0 .25 M1 E H A I 0 .25 M2 0.5... might focus on short-term benefits (e.g., net income this year), medium-term benefits (e.g., economic and environmental risk reduction), or longterm benefits (e.g., soil conservation and natural pest population control) These different criteria are not always compatible and, at the present time, short-term benefits tend to dominate over long-term ones (García-Barrios and García-Barrios, 19 92) We can group... preliminary data (García-Barrios, 20 00) and is currently under trial © 20 03 by CRC Press LLC Species richness Low High In natural plant communities, species richness declines (either monotonically or with a hump) as soil fertility increases (Based on Abrams, 1995.) LER Figure 2. 19 Productivity advantage disadvantage (Rao and Willey,1980) (Natarajan and Willey,1986) 1.0 (García-Barrios, 20 00) LER Low advantage . w 1 2 3 5 4 6 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Y 1 2 3 5 4 6 0 number of sp. A plants per unit area (a) (b) 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 0 0.1 0 .2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 Six. weight gain does not fully com- pensate species 1’s per-plant loss. M1 A BC D F E G H I 0 .25 M2 0.5 M2 0.75 M2 M2 Additive designs Substitutive designs LER =2 LER=1 Crop 2 yield per unit area Crop1. details. w 6 5 4 1 2 3 0 Y 6 5 4 1 2 3 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Number of sp. A plants per unit area Six possible scenarios: 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 1: